a'+ Bis(tetramethylhomocyclopropeny1) Dication - American Chemical

configuration assigned to 13 was confirmed using Eu(fod)3. Although there is no assurance that exo- la and 12 react ac- cording to the same mechanism,...
7 downloads 0 Views 376KB Size
-

and much less exothermic than the allylic analogue 9 10, the former reaction is far slower ( t l / 2 5 5 0 = 0.5 h). Rearrangement of 12 was found to be stereospecific, and the exo configuration assigned to 13 was confirmed using Eu(fod)3. Although there is no assurance that exo- la and 12 react according to the same mechanism, it is striking that the rearrangement stereochemistry demonstrated for 12 matches that postulated for exo- la." W e propose the name pseudopericyclic to describe the [ 1,3]sigmatropic pathway shown in 7. A pseudopericyclic reaction is a concerted transformation whose primary changes in bonding compass a cyclic array of atoms, at one (or more) of which nonbonding and bonding atomic orbitals interchange roles.'8s'9 In a crucial sense, the role interchange means a "disconnection" in the cyclic array of overlapping orbitals because the atomic orbitals switching functions are mutually orthogonal. Hence pseudopericyclic reactions cannot be orbital symmetry forbidden. The mechanistic idea described here is certainly not new, but we are unaware of a clear and general statement of this concept in the literature. Many apparently pseudopericyclic reactions are known.20 Prototropy in internally hydrogen bonded enols of P-dicarbonyl compounds (eq 1) is a prosaic example. As the proton tunnels between minima,

4327 (6) The Raman spectrum was measured by Dr. Charles W. Gillies at Harvard University. (7) This band is split slightly (8-9 cm-') on the low frequency side, perhaps as a consequence of Fermi resonance with the band at 831 cm-'. Incidentally, the vibrational spectra also rule out a rectangularlydistorted ( C z v ) 1b. (8) Professor C. Hackett Bushweller of Worcester Polytechnic Institutecarried out this experiment. (9) Y. Kobayashi, I. Kumadaki. A. Ohsawa, and Y. Sekine, Tetrahedron Len., 2841 (1974). IO) Though reported as an this diene has been obtained as white crystals, mp 37-36': ir (vapor) 1724, 1344, 1276, 1218, 1183, 1163, 714 cm-'; 19FNMR (CDC13)6 17.40 (m), 15.43 (m); 'H NMR (CDCI3)6 6.62 (m), 4.86 (m); MS 392 (parent). 11) (a) R. 5.Woodward and R. Hoffmann. "The Conservation of Orbital Symmetry", Academic Press, New York. N.Y., 1970; (b) J. A. krson, T. Miyashi, and G. Jones, 11, J. Am. Chem. SOC., 96,3468 (1974), and references contained therein. 12) Even though ring opening in cis- and trans-2.3-diphenylthiirane oxide produces a biradical with benzylic stabilization, these compounds cleave only slowly at room temperature (K. Kondo. M. Matsumoto, and A. Negishi, Tetrahedron Lek, 2131 (1972)). 13) Suitable geometry is a necessary, but not a sufficient criterion for a very low barrier. The Dewar thiophene (2) rearranges perceptibly by NMR only above 100 OC. Attempts to oxidize exo-la to the sulfone, which of course lacks the lone pair on sulfur, have met with failure. 14) F. Lautenschlaeger, J. Org. Chem., 34, 3998 (1969). 15) Sufficient "proton sponge" was present in some runs to neutralize the trifluoroacetic acid formed, thus assuring that acid catalysis was not responsible for the rapid rate. 16) P. Chao and D. M. Lemal, J. Am. Chem. SOC., 95,920 (1973), and references contained therein. 17) Oxidation of i with periodate or mchloroperbenzoic acid (-30 to 0') gives I/.With the reasonable assumption that oxidation occurs on the ex0 face of i, this stereochemical result parallels those under discussion. Sulfoxide li is capable of further rearrangement on heating, but by a different mechanism. A. G. Anastassiou. J. C. Wetzel, and B. Y.-H. Chao. J. Amer. Chem. SOC.,97, 1124(1975).

lone pair and bonding orbitals formally interchange functions a t both oxygens in the planar chelate ring.2'-23 In olefin hydroboration (eq 2) the vacant boron orbital presumably

0

0 -

I

(18) A bonding-nonbonding interchange involving different atoms does not qualify; e.g.. lil iv., a pericyclic reaction:

P-+

SR

"SR 111

switches roles with the orbital employed in bonding to hydrogen. By this device a planar, four-center transition state, normally very high lying, becomes easily a c c e s ~ i b l e . ~ ~ Further studies of sulfoxide rearrangement mechanisms are in progress. Acknowledgment. The authors wish to give special thanks to Professor C. Hackett Bushweller and Dr. Charles W. Gillies for making the spectroscopic measurements cited, and to Professors Orville Chapman, Roald Hoffmann, and R. B. Woodward for valuable criticism. They are grateful as well to the National Science Foundation and the donors of the Petroleum Research Fund, administered by the American Chemical Society, for generous financial support. References and Notes (1) H. A. Wiebe, S. Braslavsky. and J. Heicklen. Can. J. Chem., 50, 2721 (1972). (2) All "F NMR chemical shifts are expressed in parts per million downfield from external trifluoroacetic acid. (3) A square pyramidal structure has also been postulated by R. Hoffmann and W. D. Stohrer for (CH),' and its derivatives. See H. Hogeveen and P. W. Kwant, ACC. Chem. Res., 8, 413 (1975), and references contained therein. (4) In the ground state of SO there is an orthogonal pair of singly occupied IC* orbitals of the proper symmetry to mix with & and 4b3 of the diene, which are singly occupied in the Huckei approximation. Moreover, the corresponding pair of filled T orbitals in SO could transfer charge to the diene by mixing with $? and $3, while vacant d orbitals provide for backdonation to sulfur from $,, $2, and $3. R. Hoffmann emphasizes, however, that the sulfur lone pair-$' interaction would be strongly repulsive. (5) In local D,, symmetry this transition is rigorously forbidden. While the axial perturbation should confer some intensity on this Raman band, it is expected to be weak. Similarly. the ring breathing mode (ir forbidden in local D4,, symmetry) of the tetramethylcyclobutadiene-nickel chloride complex gives rise to a weak band in the infrared (H. P. Fritz in Adv. Organomet. Chem., 1, 258 (1964).

(19) (20) (21)

(22) (23)

(24)

a ' + iv

In cheletropic reactions, two atomic orbitals at the same center are intimately involved, but they do not interchange roles. For the definition of pericyclic reaction, see R. B. Woodward and R. Hoffmann, ref 1l a . Numerous examples will be presented in a full paper. A bonding p orbital and a "nonbonding" lone pair orbital switch roles at the left-hand oxygen, while a complementary interchange occurs at the oxygen on the right. Although the bondinglnonbonding distinction is obviously not absolute here, the separation of this concerted process into persistently orthogonal u and P components dramatizes the fact that it is not pericyclic. The crystal structure of dibenzoylmethane, as its enol, is presented in R. W. G. Wyckoff, "Crystal Structures", 2d ed, Vol. 6, Interscience, New York, N.Y., 1971, pp 125-128. Many other prototropic processes in conjugated systems are presumably pseudopericyclic, among them the proton tunneling phenomena in DNA which Lowdin has proposed as a basis for mutagenesis (W. C. Hamilton and J. A. Ibers, "Hydrogen Bonding in Solids", W. A. Benjamin, New York, N.Y., 1968, pp 176-178). K. Fukui, Bull. Chem. Soc. Jpn., 39, 498 (1966).

James A. Ross, Reginald P. Seiders, David M. Lemal* Department of Chemistry, Dartmouth College Hanouer, New Hampshire 03755 Receiued February 10, 1976 Novel Aromatic Systems. 6.' The Bis(tetramethylhomocyclopropeny1) Dication

Sir: In view of our current interest in the two T electron homoaromatic cyclobutenyl(homocyclopropeny1) cation 1,2we wish to report now the interesting and unexpected finding that the two electron oxidative ring opening reaction of the tetramethylcyclobutadiene dimer 23 in antimony pentafluoCommunications to the Editor

4328 ride-sulfuryl chloride fluoride solution gives the bis(tetramethylhomocyclopropenyl) dication 3.

P

Table I.

IH and I3C NMR Parametersn

Ion 1

1,3

2

7.95 9.72 133.5 187.6 6IH 6 1 3 ~ 179.8 186.2 61~ 613C 237.1 158.0 6 1 ~ 11.3 dl3C 230.6 164.5 ~ I H

613C

3

2, R = CH, 3, R = CH, 4,R=H 5,R= H We previously have observed the exclusive formation of the octamethyldihydropentalene dication 6 from isomeric syn- and ~nti-octamethyltricyclo[4.2.0.0~~~]octadienes (2), octamethylcyclooctatetraene (7), and octamethylsemibullvalene (8) in fluorosulfuric acid-antimony pentafluoride-sulfur dioxide s o l ~ t i o nWe . ~ recently have also found that antimony pentafluoride oxidizes 1,3,5,7-tetramethylcyclooctatetraene (10) to the 8 C - 6 ~cyclooctetraene dication 12, which upon

6, R, = R, = CH,

Rl

R1

7, R, = RL= C H I

9, R, = H;R, = CH,

6

9 13

~ I H 6l3c

14

--

171.3

61~ 613c

195.1

-

171.3 -

174.4

Skeletal position CI-Me C2-Me C4-Me A6i3cb

4 4.53 54.0 -

81.0 -

84.2 88.3 4.53 57.8

-_ 76.0

-

2.64 14.0 3.29 25.6 -

2.63 13.5 2.91 14.0

-

-

-

-

2.58 15.1 2.30 21.8 3.40 20.8 2.37 12.9 2.43 9.80

1.68 11.4 2.05 11.4 3.02 14.6 1.39 11.1 2.38 27.3

-45.9 -6.4 79.1 66.1

0.0 17.1

Proton and carbon-13 chemical shifts are given in parts per million from external Me& (capillary). L613C = 6cl,3- ~ C Z . tenyl cations, and we have recently reported both proton and carbon-13 N M R spectra of a series of these ions. Both the proton and carbon-13 N M R spectra of the species obtained from syn- 2 are quite similar to those of the tetramethyl- and tetramethylchlorocyclobutenyl cations, 13 and 14 (Table I), but show more deshielding of C2 and C2'. The N M R data clearly indicate structure 3 for the species obtained from 2 in SbF5-SO2CIF solution.

10. R, = H;RL= CH,

11, R,= R,= H

CH3 13.R = H 14. R = CI

8, R = CH,

warming immediately rearranged to the tetramethyldihydropentalene dication 9.5 N o ring opening reaction from ion 12 was observed. In contrast, when syn-octamethyltricy-

10

SbF,-SO.CIF -7a"

CHI)$

:\-/' if j

--10"9

CH >

CH $ 12

cl0[4.2.0.0~~~]octadiene ( 2 ) (as a suspension in SO2ClF solution) was carefully introduced into a well-stirred SbF5-SO2CIF solution at either -78 or - l o o , the resulting dark brown solution showed a ' H N M R spectrum completely different from that of 6.5It consisted of three singlets at 6 2.64, 2.58, and 1.68 in the ratio of 2: 1:l (relative to capillary Me&, -6OO). Within the temperature range studied (-125 to 0') only slight broadening of the two most shielded singlets was observed, while the third at 6 2.64 remained relatively sharp. The FT I3C N M R (proton-coupled) spectrum of the solution showed three singlets a t 613C 186.2, 179.8, and 81.0in theratioof 1:2:1 and three quartets centered a t 613C 15.1 (1 C), 14.0 (2 C), and 11.4 (1 C). Again, the I3C N M R spectrum obtained from syn- 2 in SbFs-SO2CIF solution was quite different from that of 6 obtained in F S O ~ H - S ~ F S - S solution. O~ The 'H and 13C N M R spectroscopic data are summarized in Table I for comparison. Katz and Gold6 first prepared several polymethylcyclobuJournal of the American Chemical Society / 98.14

/

Our previous studies showed a continuum of increasing homoaromatic contributions in going from fully substituted to the parent cyclobutenyl cation and clearly established the puckered conformation of the latter, Le., the homocyclopropenyl cation 1.2 Substantial 1,3-orbital interaction, intermediate between u and R, drastically reverses the order of deshielding between the terminal ( C I and C3) and central (C2) allylic carbon^.^,' The terminal carbons in the allylic fragment of ion 3 are found to be 6.4 ppm less deshielded than is the central carbon. The difference in carbon shifts (asl3,) for ions 13 and 14 are 0.0 and 17.1 ppm, respectively. Positive charge is found to be shared predominantly by the terminal allylic carbons in dihydropentalene dications 6 (Ad13c = 79.1) and 9 (A613c = 66.1). Steric constraint might be the reason of better 1,3-overlap in ion 3 as indicated by the higher deshielding a t the central allylic carbon over the terminal ones, thus causing 3 to be a truly homoaromatic dication. The bis(tetramethylhomocyclopropeny1) dication 3 apparently is quite stable and shows no tendency to become transformed into the dihydropentalene dication 6. Attempts to generate the parent bishomocyclopropenyl dication 5 from the cyclobutadiene dimers 4 (syn and anti)s have, however, failed. This observation is also in accord with the apparent thermal instability of the parent cyclooctatetraene dication 15, the preparation of which has to date remained elusive. Acknowledgment. Support of our work by the National Science Foundation is gratefully acknowledged. References and Notes (1)Part 5. G.A. Olah and G. Liang, J. Am. Chem. Soc., 98,3033 (1976). (2)(a) G.A. Olah. J. S.Staral, and G. Liang, J. Am. Chem. Soc., 96,6233(1974); (b) G.A. Olah, J. S. Staral, R. J. Spear, and G. Liang, ibid., 97,5489 (1975). (3)H. M. Rosenberg and E. C. Eimutis, Can. J. Chem., 45, 2263 (1967);R . Criegee and G. Louis, Chem. Ber., 90,417 (1957).

July 7 , 1976

4329 (4) J. M. Bollinger and G. A. Olah. J. Am. Chem. SOC.,91, 3380 (1969). (5) G. A. Olah, J. S. Staral, and L. A. Paquette. J. Am. Chem. SOC.,98, 1267 (1976). (6) T. J. Katz and E. H. Gold, J. Am. Chem. SOC.,86, 1600 (1964). (7) G. A. Olah and G. Liang. J. Am. Chem. SOC., 94, 6434 (1972); 97, 1987 (1975). (8) M. Avram, I. G. Dinulescu,E. Marcia, G. Mateescu, E. Sliam, and C. D. Nenitzescu, Chem. &., 97,382 (1964). The authors thank Michael J. Carmody for samples of these hydrocarbons.

George A. Olah,* Gao Liang Department of Chemistry, Case Western Reserve University Cleveland. Ohio 44106 Leo A. Paquette,* William P. Melega Department of Chemistry, The Ohio State Unicersity Columbus, Ohio 43210 Received April 13, 1976

A Trans Cyclohexene

0.m . . . ., , 0

1

,

,

.,. . . _,.. . . , _. . . 2

3

4

.. .

5

.

,

6

[PIx dt.4 . . ..I.. . .,. . . .,. . . 7

8

9

-

Figure 1. Dependence of the quantum yields 4 and 4' on the hydrogen ion concentration, I H+I. The curves are the theoretical variations of 4 (solid line) and 4' (dotted line) calculated using the expression (eq 2) given in the text. Plots are experimental values of 4 measured by uv spectroscopy (0) and GC (A)and of 4' measured by gc (e).

Sir:

We wish to report experimental evidence for a twisted form of 1-phenylcyclohexene presenting a double bond past orthogonality and commonly called trans- 1-phenylcyclohexene. Laser photolysis1 of 1-phenylcyclohexene (I) in methanol solution a t room temperature forms a transient species which absorbs in the range 300-430 nm with a maximum near 380 nm. The lifetime of this transient, 7 = l / k l , is 9 ps and it is unaffected by the presence of dissolved oxygen. A transient species with absorption characteristics similar to those obtained in methanol solution is also observed in the laser photolysis of I in acetonitrile and cyclohexane solution (lifetimes: 14 and 9 KS, respectively). In methanol solution, the 1-phenylcyclohexene transient is quenched by hydrogen ion with a rate constant kQ = 7.6 X lo6 M-I s-I, as found by plotting the reciprocal of the decay-time of the transient as a function of added acid concentration. This sensitivity of the transient species to proton concentration is a n indication that it may be identical with the intermediate postulated by Kropp2 on purely photochemical grounds. W e now proceed to demonstrate this identity. Let us consider the classical mechanism of acid-catalyzed photodddition of methanol, where X is the reactive intermediate and I1 the methyl 1 -phenylcyclohexyl ether: hu

1-x

x-I X

+ H+

--

p (quantum yield)

ko

carbocation

carbocation

MeOH

I1

k d H+l a (chemical yield)

completed by reactions accounting for photochemical disappearance of I in neat methanol.

X

+ MeOH MeOH.1

X

-

carbocation

+ MeO-

products

k' I MeOHl = k', k,"

If 4 and 4' are respectively the quantum yields of disappearance of I and of formation of 11, then:

and

+

with kl = ko + k,' + k,", k , = k,' kr"and p' = ap. The value of 4 for any H+ concentration is given by the relation:

From two particular values of 4, measured in neat methanol and in an acid solution ( H + = 0.096 N) and respectively denoted &,and 41, combined with the values of kl and kQ measured in the laser-photolysis experiments reported above, we calculate: first, the value of cp = 0.082 obtained from

derived from relation 2 then the function

represented in Figure 1 (solid line). Relations 1 and 1' having identical functional dependence, the same treatment applies to quantum yields d', do', $l', and p'. Measurements of 4 and 4' for other H + concentrations give the experimental values also plotted in Figure 1. The good agreement of the experimental plots with the calculated curves indicates that relation 2 is verified within experimental errors. Therefore the ratio k q / k l is the same for the reactive intermediate involved in the above mechanism and for the transient observed in laser-photolysis experiments. Such a coincidence would be highly improbable if there were two different species; this leads us to conclude that the transient observed in laser photolysis is the reactive intermediate involved in Markovnikov addition reactions. The observed intermediate can be potentially identified with (a) orthogonal triplet phenylcyclohexene A, (b) an orthogonal zwitterionic phenylcyclohexene B, or (c) the trans phenylcyclohexene C (with partial singlet diradical character if the twist is not fully completed to 180'). Rosenberg and Serve5 identified the reactive intermediate with the triplet state. In fact their experiments, as well as many other works6 on sensitization of the photosolvatation of cycloalkenes by triplet donors, demonstrate only that the triplet state is either the reactive intermediate or a precursor of this intermediate. The intermediate is not quenched by dissolved oxygen and its lifetime would be surprisingly long for an orthogonal alkene triplet;' hence intermediate A is highly improbable but our work is consistent with the possibility that the triplet, like the singlet, is a precursor of the intermediate, without being the intermediate itself. The chemistry of the intermediate is consistent with both B and C. Although our initial purpose was to trap a zwitterionic intermediate such as B, the relative insensitivity of lifetime to the polar nature of the solvent appears to rule out the solCommunications to the Editor